Hybrid LSTM-Transformer Models: A Breakthrough in OWC Device Power Generation Prediction

This research study has reviewed the power gener-ation prediction problem associated with the oscillating water column(OWC) wave energy converter device and proposed a hybrid LSTM- Transformer layer model to predict the power generated by the OWC device. OWC front wall draft, OWC duct length, Chamber length, ripple number of seabed and ripple amplitude of seabed. While the response variable is the power generated by the OWC device. The input samples are taken as five equally spaced points in the specified range of the inputs. A there are five input attributes and five unique values for each attribute, a set of 3125 examples are there to train the model. The response variable for each example is calculated by the numerical model based on boundary element method. It is obtained that the hybrid LSTM-Transformer model is able to accuratly predict the power generated by the OWC device with an R2 value of 0.99963 for the test set.